#include "tree.h"
/*****************************************************************/
-Tree::Tree(int num) {
+Tree::Tree(int num, TreeMap* t) : tmap(t) {
try {
- globaldata = GlobalData::getInstance();
m = MothurOut::getInstance();
numLeaves = num;
}
}
/*****************************************************************/
-Tree::Tree(string g) {
+Tree::Tree(string g) { //do not use tree generated by this its just to extract the treenames, its a chicken before the egg thing that needs to be revisited.
try {
- globaldata = GlobalData::getInstance();
m = MothurOut::getInstance();
- parseTreeFile(); globaldata->runParse = false;
+ tmap = NULL;
+
+ parseTreeFile(); m->runParse = false;
}
catch(exception& e) {
m->errorOut(e, "Tree", "Tree - just parse");
}
}
/*****************************************************************/
-Tree::Tree() {
+Tree::Tree(TreeMap* t) : tmap(t) {
try {
- globaldata = GlobalData::getInstance();
m = MothurOut::getInstance();
- if (globaldata->runParse == true) { parseTreeFile(); globaldata->runParse = false; }
+ if (m->runParse == true) { parseTreeFile(); m->runParse = false; }
//for(int i = 0; i < globaldata->Treenames.size(); i++) { cout << i << '\t' << globaldata->Treenames[i] << endl; }
- numLeaves = globaldata->Treenames.size();
+ numLeaves = m->Treenames.size();
numNodes = 2*numLeaves - 1;
tree.resize(numNodes);
-
+
//initialize groupNodeInfo
- for (int i = 0; i < globaldata->gTreemap->namesOfGroups.size(); i++) {
- groupNodeInfo[globaldata->gTreemap->namesOfGroups[i]].resize(0);
+ for (int i = 0; i < (tmap->getNamesOfGroups()).size(); i++) {
+ groupNodeInfo[(tmap->getNamesOfGroups())[i]].resize(0);
}
-
+
//initialize tree with correct number of nodes, name and group info.
for (int i = 0; i < numNodes; i++) {
//initialize leaf nodes
if (i <= (numLeaves-1)) {
- tree[i].setName(globaldata->Treenames[i]);
+ tree[i].setName(m->Treenames[i]);
//save group info
- string group = globaldata->gTreemap->getGroup(globaldata->Treenames[i]);
+ string group = tmap->getGroup(m->Treenames[i]);
+
vector<string> tempGroups; tempGroups.push_back(group);
tree[i].setGroup(tempGroups);
groupNodeInfo[group].push_back(i);
tree[i].pGroups[group] = 1;
//Treemap knows name, group and index to speed up search
- globaldata->gTreemap->setIndex(globaldata->Treenames[i], i);
+ tmap->setIndex(m->Treenames[i], i);
//intialize non leaf nodes
}else if (i > (numLeaves-1)) {
tree[i].setGroup(tempGroups);
}
}
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "Tree", "Tree");
+ exit(1);
+ }
+}
+/*****************************************************************/
+Tree::Tree(TreeMap* t, vector< vector<double> >& sims) : tmap(t) {
+ try {
+ m = MothurOut::getInstance();
+
+ if (m->runParse == true) { parseTreeFile(); m->runParse = false; }
+ //for(int i = 0; i < globaldata->Treenames.size(); i++) { cout << i << '\t' << globaldata->Treenames[i] << endl; }
+ numLeaves = m->Treenames.size();
+ numNodes = 2*numLeaves - 1;
+
+ tree.resize(numNodes);
+
+ //initialize groupNodeInfo
+ for (int i = 0; i < (tmap->getNamesOfGroups()).size(); i++) {
+ groupNodeInfo[(tmap->getNamesOfGroups())[i]].resize(0);
+ }
+
+ //initialize tree with correct number of nodes, name and group info.
+ for (int i = 0; i < numNodes; i++) {
+ //initialize leaf nodes
+ if (i <= (numLeaves-1)) {
+ tree[i].setName(m->Treenames[i]);
+
+ //save group info
+ string group = tmap->getGroup(m->Treenames[i]);
+
+ vector<string> tempGroups; tempGroups.push_back(group);
+ tree[i].setGroup(tempGroups);
+ groupNodeInfo[group].push_back(i);
+
+ //set pcount and pGroup for groupname to 1.
+ tree[i].pcount[group] = 1;
+ tree[i].pGroups[group] = 1;
+
+ //Treemap knows name, group and index to speed up search
+ tmap->setIndex(m->Treenames[i], i);
+
+ //intialize non leaf nodes
+ }else if (i > (numLeaves-1)) {
+ tree[i].setName("");
+ vector<string> tempGroups;
+ tree[i].setGroup(tempGroups);
+ }
+ }
+
+ //build tree from matrix
+ //initialize indexes
+ map<int, int> indexes; //maps row in simMatrix to vector index in the tree
+ int numGroups = (tmap->getNamesOfGroups()).size();
+ for (int g = 0; g < numGroups; g++) { indexes[g] = g; }
+
+ //do merges and create tree structure by setting parents and children
+ //there are numGroups - 1 merges to do
+ for (int i = 0; i < (numGroups - 1); i++) {
+ float largest = -1000.0;
+
+ if (m->control_pressed) { break; }
+
+ int row, column;
+ //find largest value in sims matrix by searching lower triangle
+ for (int j = 1; j < sims.size(); j++) {
+ for (int k = 0; k < j; k++) {
+ if (sims[j][k] > largest) { largest = sims[j][k]; row = j; column = k; }
+ }
+ }
+
+ //set non-leaf node info and update leaves to know their parents
+ //non-leaf
+ tree[numGroups + i].setChildren(indexes[row], indexes[column]);
+
+ //parents
+ tree[indexes[row]].setParent(numGroups + i);
+ tree[indexes[column]].setParent(numGroups + i);
+
+ //blength = distance / 2;
+ float blength = ((1.0 - largest) / 2);
+
+ //branchlengths
+ tree[indexes[row]].setBranchLength(blength - tree[indexes[row]].getLengthToLeaves());
+ tree[indexes[column]].setBranchLength(blength - tree[indexes[column]].getLengthToLeaves());
+
+ //set your length to leaves to your childs length plus branchlength
+ tree[numGroups + i].setLengthToLeaves(tree[indexes[row]].getLengthToLeaves() + tree[indexes[row]].getBranchLength());
+
+
+ //update index
+ indexes[row] = numGroups+i;
+ indexes[column] = numGroups+i;
+
+ //remove highest value that caused the merge.
+ sims[row][column] = -1000.0;
+ sims[column][row] = -1000.0;
+
+ //merge values in simsMatrix
+ for (int n = 0; n < sims.size(); n++) {
+ //row becomes merge of 2 groups
+ sims[row][n] = (sims[row][n] + sims[column][n]) / 2;
+ sims[n][row] = sims[row][n];
+ //delete column
+ sims[column][n] = -1000.0;
+ sims[n][column] = -1000.0;
+ }
+ }
+
+ //adjust tree to make sure root to tip length is .5
+ int root = findRoot();
+ tree[root].setBranchLength((0.5 - tree[root].getLengthToLeaves()));
}
catch(exception& e) {
m->errorOut(e, "Tree", "Tree");
exit(1);
}
}
-
/*****************************************************************/
Tree::~Tree() {}
/*****************************************************************/
-void Tree::addNamesToCounts() {
+void Tree::addNamesToCounts(map<string, string> nameMap) {
try {
//ex. seq1 seq2,seq3,se4
// seq1 = pasture
string name = tree[i].getName();
- map<string, string>::iterator itNames = globaldata->names.find(name);
+ map<string, string>::iterator itNames = nameMap.find(name);
- if (itNames == globaldata->names.end()) { m->mothurOut(name + " is not in your name file, please correct."); m->mothurOutEndLine(); exit(1); }
+ if (itNames == nameMap.end()) { m->mothurOut(name + " is not in your name file, please correct."); m->mothurOutEndLine(); exit(1); }
else {
vector<string> dupNames;
- m->splitAtComma(globaldata->names[name], dupNames);
+ m->splitAtComma(nameMap[name], dupNames);
map<string, int>::iterator itCounts;
int maxPars = 1;
set<string> groupsAddedForThisNode;
for (int j = 0; j < dupNames.size(); j++) {
- string group = globaldata->gTreemap->getGroup(dupNames[j]);
+ string group = tmap->getGroup(dupNames[j]);
if (dupNames[j] != name) {//you already added yourself in the constructor
try {
//Treemap knows name, group and index to speed up search
// getIndex function will return the vector index or -1 if seq is not found.
- int index = globaldata->gTreemap->getIndex(searchName);
+ int index = tmap->getIndex(searchName);
return index;
}
void Tree::setIndex(string searchName, int index) {
try {
//set index in treemap
- globaldata->gTreemap->setIndex(searchName, index);
+ tmap->setIndex(searchName, index);
}
catch(exception& e) {
m->errorOut(e, "Tree", "setIndex");
//float A = clock();
//if user has given a names file we want to include that info in the pgroups and pcount info.
- if(globaldata->names.size() != 0) { addNamesToCounts(); }
+ if(m->names.size() != 0) { addNamesToCounts(m->names); }
//build the pGroups in non leaf nodes to be used in the parsimony calcs.
for (int i = numLeaves; i < numNodes; i++) {
}
}
/*****************************************************************/
-void Tree::getSubTree(Tree* copy, vector<string> Groups) {
+//assumes leaf node names are in groups and no names file - used by indicator command
+void Tree::getSubTree(Tree* Ctree, vector<string> Groups) {
try {
-
+
+ //copy Tree since we are going to destroy it
+ Tree* copy = new Tree(tmap);
+ copy->getCopy(Ctree);
+ copy->assembleTree("nonames");
+
//we want to select some of the leaf nodes to create the output tree
//go through the input Tree starting at parents of leaves
for (int i = 0; i < numNodes; i++) {
tree[i].setName(Groups[i]);
//save group info
- string group = globaldata->gTreemap->getGroup(Groups[i]);
+ string group = tmap->getGroup(Groups[i]);
vector<string> tempGroups; tempGroups.push_back(group);
tree[i].setGroup(tempGroups);
groupNodeInfo[group].push_back(i);
tree[i].pGroups[group] = 1;
//Treemap knows name, group and index to speed up search
- globaldata->gTreemap->setIndex(Groups[i], i);
+ tmap->setIndex(Groups[i], i);
//intialize non leaf nodes
}else if (i > (numLeaves-1)) {
copy->tree[i].setParent(grandparent);
copy->tree[i].setBranchLength((copy->tree[i].getBranchLength()+copy->tree[parent].getBranchLength()));
- copy->tree[grandparent].setChildren(grandparentLC, grandparentRC);
+ if (grandparent != -1) {
+ copy->tree[grandparent].setChildren(grandparentLC, grandparentRC);
+ }
removedLeaves.insert(sibIndex);
}
}else{
copy->tree[sibIndex].setParent(grandparent);
copy->tree[sibIndex].setBranchLength((copy->tree[sibIndex].getBranchLength()+copy->tree[parent].getBranchLength()));
- copy->tree[grandparent].setChildren(grandparentLC, grandparentRC);
+ if (grandparent != -1) {
+ copy->tree[grandparent].setChildren(grandparentLC, grandparentRC);
+ }
removedLeaves.insert(i);
}else{
//neither of us are, so we want to eliminate ourselves and our parent
int parentsSibIndex;
if (grandparent != -1) {
int greatgrandparent = copy->tree[grandparent].getParent();
- int greatgrandparentLC = copy->tree[greatgrandparent].getLChild();
- int greatgrandparentRC = copy->tree[greatgrandparent].getRChild();
+ int greatgrandparentLC, greatgrandparentRC;
+ if (greatgrandparent != -1) {
+ greatgrandparentLC = copy->tree[greatgrandparent].getLChild();
+ greatgrandparentRC = copy->tree[greatgrandparent].getRChild();
+ }
int grandparentLC = copy->tree[grandparent].getLChild();
int grandparentRC = copy->tree[grandparent].getRChild();
copy->tree[parentsSibIndex].setParent(greatgrandparent);
copy->tree[parentsSibIndex].setBranchLength((copy->tree[parentsSibIndex].getBranchLength()+copy->tree[grandparent].getBranchLength()));
- copy->tree[greatgrandparent].setChildren(greatgrandparentLC, greatgrandparentRC);
+ if (greatgrandparent != -1) {
+ copy->tree[greatgrandparent].setChildren(greatgrandparentLC, greatgrandparentRC);
+ }
}else{
- copy->tree[parent].setChildren(-1, -1);
- cout << "issues with making subtree" << endl;
+ copy->tree[parent].setParent(-1);
+ //cout << "issues with making subtree" << endl;
}
removedLeaves.insert(sibIndex);
removedLeaves.insert(i);
//you found the root
if (copy->tree[i].getParent() == -1) { root = i; break; }
}
-
+
int nextSpot = numLeaves;
populateNewTree(copy->tree, root, nextSpot);
-
+
+ delete copy;
}
catch(exception& e) {
- m->errorOut(e, "Tree", "getCopy");
+ m->errorOut(e, "Tree", "getSubTree");
+ exit(1);
+ }
+}
+/*****************************************************************/
+//assumes nameMap contains unique names as key or is empty.
+//assumes numLeaves defined in tree constructor equals size of seqsToInclude and seqsToInclude only contains unique seqs.
+int Tree::getSubTree(Tree* copy, vector<string> seqsToInclude, map<string, string> nameMap) {
+ try {
+
+ if (numLeaves != seqsToInclude.size()) { m->mothurOut("[ERROR]: numLeaves does not equal numUniques, cannot create subtree.\n"); m->control_pressed = true; return 0; }
+
+ getSubTree(copy, seqsToInclude);
+ if (nameMap.size() != 0) { addNamesToCounts(nameMap); }
+
+ //build the pGroups in non leaf nodes to be used in the parsimony calcs.
+ for (int i = numLeaves; i < numNodes; i++) {
+ if (m->control_pressed) { return 1; }
+
+ tree[i].pGroups = (mergeGroups(i));
+ tree[i].pcount = (mergeGcounts(i));
+ }
+
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "Tree", "getSubTree");
exit(1);
}
}
return (index++);
}else { //you are a leaf
- int indexInNewTree = globaldata->gTreemap->getIndex(oldtree[node].getName());
+ int indexInNewTree = tmap->getIndex(oldtree[node].getName());
return indexInNewTree;
}
}
}
}
/**************************************************************************************************/
-
void Tree::randomLabels(vector<string> g) {
try {
//initialize groupNodeInfo
- for (int i = 0; i < globaldata->gTreemap->namesOfGroups.size(); i++) {
- groupNodeInfo[globaldata->gTreemap->namesOfGroups[i]].resize(0);
+ for (int i = 0; i < (tmap->getNamesOfGroups()).size(); i++) {
+ groupNodeInfo[(tmap->getNamesOfGroups())[i]].resize(0);
}
for(int i = 0; i < numLeaves; i++){
exit(1);
}
}
-/**************************************************************************************************
-
-void Tree::randomLabels(string groupA, string groupB) {
- try {
- int numSeqsA = globaldata->gTreemap->seqsPerGroup[groupA];
- int numSeqsB = globaldata->gTreemap->seqsPerGroup[groupB];
-
- vector<string> randomGroups(numSeqsA+numSeqsB, groupA);
- for(int i=numSeqsA;i<randomGroups.size();i++){
- randomGroups[i] = groupB;
- }
- random_shuffle(randomGroups.begin(), randomGroups.end());
-
- int randomCounter = 0;
- for(int i=0;i<numLeaves;i++){
- if(tree[i].getGroup() == groupA || tree[i].getGroup() == groupB){
- tree[i].setGroup(randomGroups[randomCounter]);
- tree[i].pcount.clear();
- tree[i].pcount[randomGroups[randomCounter]] = 1;
- tree[i].pGroups.clear();
- tree[i].pGroups[randomGroups[randomCounter]] = 1;
- randomCounter++;
- }
- }
- }
- catch(exception& e) {
- m->errorOut(e, "Tree", "randomLabels");
- exit(1);
- }
-}
/**************************************************************************************************/
void Tree::randomBlengths() {
try {
void Tree::createNewickFile(string f) {
try {
int root = findRoot();
- //filename = m->getRootName(globaldata->getTreeFile()) + "newick";
+
filename = f;
m->openOutputFile(filename, out);
if (tree[node].getBranchLength() != -1) {
out << ":" << tree[node].getBranchLength();
}
+ }else if (mode == "deunique") {
+ //if there is a branch length then print it
+ if (tree[node].getBranchLength() != -1) {
+ out << ":" << tree[node].getBranchLength();
+ }
}
}else { //you are a leaf
- string leafGroup = globaldata->gTreemap->getGroup(tree[node].getName());
+ string leafGroup = tmap->getGroup(tree[node].getName());
if (mode == "branch") {
out << leafGroup;
if (tree[node].getBranchLength() != -1) {
out << ":" << tree[node].getBranchLength();
}
+ }else if (mode == "deunique") {
+ map<string, string>::iterator itNames = m->names.find(tree[node].getName());
+
+ string outputString = "";
+ if (itNames != m->names.end()) {
+
+ vector<string> dupNames;
+ m->splitAtComma((itNames->second), dupNames);
+
+ if (dupNames.size() == 1) {
+ outputString += tree[node].getName();
+ if (tree[node].getBranchLength() != -1) {
+ outputString += ":" + toString(tree[node].getBranchLength());
+ }
+ }else {
+ outputString += "(";
+
+ for (int u = 0; u < dupNames.size()-1; u++) {
+ outputString += dupNames[u];
+
+ if (tree[node].getBranchLength() != -1) {
+ outputString += ":" + toString(0.0);
+ }
+ outputString += ",";
+ }
+
+ outputString += dupNames[dupNames.size()-1];
+ if (tree[node].getBranchLength() != -1) {
+ outputString += ":" + toString(0.0);
+ }
+
+ outputString += ")";
+ if (tree[node].getBranchLength() != -1) {
+ outputString += ":" + toString(tree[node].getBranchLength());
+ }
+ }
+ }else {
+ outputString = tree[node].getName();
+ //if there is a branch length then print it
+ if (tree[node].getBranchLength() != -1) {
+ outputString += ":" + toString(tree[node].getBranchLength());
+ }
+
+ m->mothurOut("[ERROR]: " + tree[node].getName() + " is not in your namefile, please correct."); m->mothurOutEndLine();
+ }
+
+ out << outputString;
}
}
}
}
}else { //you are a leaf
- string leafGroup = globaldata->gTreemap->getGroup(theseNodes[node].getName());
+ string leafGroup = tmap->getGroup(theseNodes[node].getName());
if (mode == "branch") {
out << leafGroup;
//only takes names from the first tree and assumes that all trees use the same names.
try {
- string filename = globaldata->getTreeFile();
+ string filename = m->getTreeFile();
ifstream filehandle;
m->openInputFile(filename, filehandle);
int c, comment;
//c = , until done with translation then c = ;
h = name.substr(name.length()-1, name.length());
name.erase(name.end()-1); //erase the comma
- globaldata->Treenames.push_back(number);
+ m->Treenames.push_back(number);
}
- if(number == ";") { globaldata->Treenames.pop_back(); } //in case ';' from translation is on next line instead of next to last name
+ if(number == ";") { m->Treenames.pop_back(); } //in case ';' from translation is on next line instead of next to last name
}
}
filehandle.close();
}
//cout << "name = " << name << endl;
- globaldata->Treenames.push_back(name);
+ m->Treenames.push_back(name);
filehandle.putback(c);
//k = c;
//cout << " after putback" << k << endl;
projects / mothur.git / blobdiff